Wang, Zhen

241 Background: In a rapidly moving field, such as cancer immunotherapy, where immune checkpoint inhibitors (ICIs) are used across 14 different tumor types, patients may receive suboptimal treatment or even be harmed if information on toxicity is not readily translated for use in clinical practice. Every single systematic review and meta-analysis which attempted to summarize toxicity of immune checkpoint inhibitors (ICIs) quickly became outdated. A living systematic review, which is defined as a systematic review that is continually updated to incorporate relevant new evidence as it becomes available, is necessary in this situation. Methods: The process of creating a living systematic review started with the creation of a comprehensive search designed by a librarian experienced in systematic reviews in collaboration with the study’s principle investigator. Search was constantly updated every 3 months and evidence is synthesized in a series of steps (microtasks) using a combination of human and augmented intelligence. A complete infrastructure is being developed and it includes automated cumulative meta-analysis and an online reporting platform which will constantly update information for clinicians and patients in a live manner. Results: We screened 6746 studies during Sep 2018-March 2019 and identified 6746 studies and we were able to successfully maintain up-to-date toxicity estimates for immune mediated adverse events over this period while maintaining the rigor of a conventional systematic review. Eventually, we will integrate the steps of LSR into one, user-friendly, semi-automated format which can independently provide accurate estimates and feed into and support a living guidelines platform through shared Application Programing Interface (APIs). Conclusions: LSRs are feasible, efficient, and when fully developed can reduce redundancy and waste in medical research, improve the quality of evidence, reduce human effort and support living and dynamic guidelines to facilitate truly informed shared decision making.

Abstract Background: Myelodysplastic syndromes (MDS) are clonal hematological diseases which present with cytopenias. Hematopoietic cell transplantation is usually limited to fit patients with higher risk MDS and donor availability. Hypomethylating agents (azacitidine and decitabine) have been the mainstay option for the management of MDS with different clinical efficacy in low versus high risk MDS trials. No trials have compared the two agents. Aim: To conduct a systematic review and network analysis comparing the efficacy of azacitidine to decitabine. Methods: The protocol of the systematic review was developed a priori. A comprehensive search of several databases (MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, and Scopus) was conducted from each database's earliest inception through November 20th, 2014 without language restrictions. Trials enrolling adults diagnosed with MDS who received hypomethylating agents (azacitidine or decitabine) therapy were included. Studies were screened by two independent reviewers and differences were resolved by consensus. The Cochrane Risk of bias tool was used to appraise the trials. Random effects model was used to pool relative risks (RR) of outcomes (overall survival, overall response rate, hematologic improvement and grade 3 or 4 toxicity). Adjusted indirect comparisons were used to estimate RR for indirect comparisons (Glenny AM, et al. Health Technol Assess. 2005). All statistical analyses were conducted using STATA, version 13 (StataCorp LP, College Station, TX). Results: Only four trials met the eligibility criteria (Figure 1). Two trials compared Azacitidine (75mg/m²/day SC x 7 days) to the best suppurative care (BSC) and included 549 patients (278 azacitidine and 271 BSC, age average: 69; range: 31-92), and the other 2 compared decitabine (15 mg/m² IV q 8 hours x9) to BSC and included 403 patients (208 Decitabine and 195 BSC, age average: 69.7; range: 60-90) (Table 1). The proportion of patients with intermediate-2 and high-risk myelodysplastic syndrome (based on International Prognostic Scoring System (IPSS)) in trials comparing decitabine to BSC was 82.21% and 83.08%; respectively, and in trials comparing azacitidine to BSC was 62.59% and 61.26%; respectively. The risk of bias was moderate overall. Compared to BSC, azacitidine was significantly associated with lower risk of death (RR=0.83, 95% CI: 0.74-0.94, p=0.002) whereas the effect of decitabine did not reach statistical significance (RR=0.88, 95% CI: 0.77-1.001, p=0.053). Both drugs were superior to BSC in terms of partial and complete response. Head to head comparisons were not statistically significant (except for the outcome of complete response where low certainty evidence suggested that azacitidine treated patients were less likely to have complete response compared to decitabine (RR=0.11, 95% CI= 0.01, 0.86, p=0.04). (Table 2). Conclusion: Azacitidine and decitabine are both superior to BSC. The available indirect evidence comparing the two agents warrants low certainty and cannot reliably confirm superiority of either agent. Head-to-head trials are needed. In the meantime, the choice of agent should be driven by patients' preferences, drug availability and cost. Table 2. Results of meta-analysis Outcome Azacitidine VS. Decitabine Azacitidine VS. BSC Decitabine VS. BSC RR LCI HCI P value RR LCI HCI P value RR LCI HCI P value Death 0.95 0.79 1.13 0.54 0.83 0.74 0.94 0.002 0.88 0.77 1.001 0.05 Complete response 0.11 0.01 0.86 0.04* 2.56 1.44 4.58 0.001 23.46 3.22 170.84 0.002 Partial response 0.35 0.04 3.03 0.34 4.91 2.27 10.63 <0.0001 14.02 1.87 105.08 0.01 Major erythroid improvement 0.47 0.03 8.37 0.60 6.37 3.93 10.33 <0.0001 13.67 0.79 235.57 0.07 Major platelet improvement 1.51 0.30 7.56 0.62 4.80 2.98 7.70 <0.0001 3.19 0.68 14.89 0.14 Major neutrophil improvement 2.89 0.56 14.82 0.20 2.63 1.68 4.12 <0.0001 0.91 0.19 4.38 0.907 Hematologic improvement 0.25 0.06 1.09 0.07 2.18 1.67 2.85 <0.0001 8.62 2.05 36.32 0.003 Anemia 1.18 0.53 2.63 0.69 0.89 0.75 1.06 0.198 0.76 0.35 1.66 0.49 Neutropenia 0.99 0.75 1.30 0.92 1.87 1.63 2.14 <0.0001 1.90 1.49 2.42 <0.0001 Thrombocytopenia 0.87 0.64 1.18 0.37 1.63 1.43 1.86 <0.0001 1.87 1.43 2.45 <0.0001 Infection 1.13 0.52 2.43 0.76 1.25 0.60 2.60 0.55 1.11 0.87 1.42 0.41 *These results warrant low certainty due to imprecision (very small number of events) and are driven by 14 patients achieving complete response in BSC arm in one trial. Disclosures Al-Kali: Celgene: Research Funding.

Abstract Context The extent to which some pharmacological interventions reduce or increase the risk of biochemical conversion to type 2 diabetes mellitus (T2DM) in at-risk individuals is unclear. Methods We searched MEDLINE, Embase, the Cochrane Central Register of Controlled Trials, the Cochrane Database of Systematic Reviews, and Scopus through 24 August 2017 for randomized controlled trials evaluating the effect of drugs suspected to modify the risk of biochemical conversion to T2DM. Results We included 43 trials with 192,156 subjects (mean age, 60 years; 56% men; mean body mass index, 30.4 kg/m2). α-Glucosidase inhibitors, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, metformin, orlistat, phentermine/topiramate, and pioglitazone significantly reduced the risk of biochemical conversion to T2DM, whereas statins and nateglinide increased the risk. There was insufficient direct evidence regarding the effects of sulfonylureas, glucagon-like peptide-1 receptor agonists, dipeptidyl peptidase-4 inhibitors, and sodium-glucose cotransporter-2 inhibitors. Most trials were brief and evaluated this outcome during treatment without a withdrawal or washout period. Conclusions Several drugs modify the risk of biochemical conversation to T2DM, although whether this effect is persistent and clinically relevant is unclear. Future studies need to focus on cardiovascular disease prevention, mortality, and patient-important outcomes instead of biochemical conversion to T2DM.

Abstract Background Regorafenib at different dosing strategies and TAS-102 are treatment options for refractory metastatic colorectal cancer (mCRC). We aimed to evaluate the comparative effectiveness evidence supporting these different strategies. Materials and Methods We searched different databases for randomized controlled trials evaluating TAS-102 or regorafenib in patients with refractory mCRC who failed prior oxaliplatin, irinotecan, and fluoropyrimidine. Outcomes of interest included overall survival (OS) and progression-free survival (PFS). The overall effect was pooled using the DerSimonian random effects model. We conducted network meta-analysis based on White's multivariate meta-regression to pool evidence from direct and indirect comparisons. Results Six trials at low risk of bias (2,445 patients) were included. Direct comparisons showed that Rego 160 and TAS-102 as monotherapy were superior to best-supportive care (BSC) in terms of PFS (Rego 160: hazard ratio [HR], 0.4; 95% confidence ratio [CI], 0.26–0.63; TAS-102: HR, 0.46 CI, 0.40–0.52) and OS (Rego 160: HR, 0.67; CI, 0.48–0.93; TAS-102: HR, 0.67; CI, 0.57–0.80). Network analysis showed no statistically difference in PFS or OS between Rego 160 and TAS-102. Rego 80+ was superior to BSC in terms of OS (HR, 0.44; CI, 0.23–0.84) and PFS (HR, 0.37; CI, 0.21–0.66). Rego 80+ was associated with statistically nonsignificant improvement in OS and PFS compared with TAS-102 and Rego 160. Conclusion Regorafenib 160 and TAS-102 appear to have similar efficacy. Rego 80+ is shown to be superior to BSC. A trend for improved OS was observed with Rego 80+ versus Rego 160 or TAS 102. Implications for Practice Regorafenib at a dose of 160 mg and TAS-102 appear to have similar efficacy in patients with refractory metastatic colorectal cancer. Regorafenib with a dose escalation strategy is superior to best-supportive care. Given its tolerability and the observed trend in survival benefit compared with regorafenib 160, dose escalation strategy of regorafenib (80+) may be the preferred option in this setting.

6596 Background: Several previous systematic reviews and meta-analyses have attempted to summarize toxicity of Immune checkpoint inhibitors (ICIs). However, very soon after each one of these reviews has been published, it became outdated. ICIs are currently used in 14 different cancers and data is rapidly evolving from new clinical trials. A living Systematic review, which is defined as a systematic review that is continually updated to incorporate relevant new evidence as it becomes available, is necessary in this situations. Therefore, we performed an updated systematic review and a meta-analysis which will serve as a foundation of a living Systematic review. Methods: MEDLINE, EMBASE and Cochrane were searched to identify phase 2 and 3 RCTs of PD-1/PD-L1 ICIs. Included studies compared either immunotherapy alone or combination with existing standard of care treatment and reported data for AE’s of interest. DerSimonian-Laird random effects Meta-Analysis was performed to derive pooled odds Ratio (OR) estimates for AE’s of interest. An infrastructure of a living systematic review is being developed and it includes monthly literature searches, cumulative meta-analysis and an online reporting platform. Results: We screened 6746 studies and 31 phase 3 and 2 phase 2 RCTs (n = 21,421) were included in the analysis. 22 RCTs used PD-1/PD-L1 ICIs as a single agent and 11 as a combination therapy. Selected toxicity estimates are summarized in a table. Conclusions: The meta-analysis updates previously published toxicity estimates and provides additional information about the risk of toxicities in single versus combination regimens. We have initiated the first living systematic review in oncology that will be continuously updated, incorporating relevant new evidence as it becomes available, and will provide accurate and up to date toxicity estimates to support clinical decision making. [Table: see text]

e12057 Background: Several RCTs have evaluated a shorter duration of T ( < 12 months) with hopes of similar efficacy, reduced cardiotoxicity, cost and burden of treatment. Of the 5 major studies, PERSPHONE showed that 6 months was non-inferior compared to 12 months. However, four additional studies of shorter duration failed to demonstrate non-inferiority. Therefore, we performed an updated systematic review and a meta-analysis to assess the optimal duration of adjuvant T. Methods: MEDLINE, EMBASE, Cochrane, Scopus and conference proceedings were searched to identify RCTs comparing one year of adjuvant T with a shorter duration in early-stage HER2+ breast cancer. The DerSimonian-Laird random-effects Meta-Analysis was performed using CMAv3 software to derive pooled Hazard Ratio (HR) estimates for overall survival (OS, Disease-Free Survival (DFS) and Odds Ratio(OR) estimates of cardiac toxicity. Q-test was used to assess between-study heterogeneity; I2 statistic was computed to express the percentage of the total observed variability due to study heterogeneity. The risk for bias was assessed using the Cochrane Collaboration’s tool. Results: We screened 1772 citations and identified 5 studies(n = 11,377) for analysis. 5691 patients were randomized to 12 months of adjuvant T while 5686 were randomized to a shorter duration (3 studies evaluated 6 vs 12 months, 2 studies evaluated 9 weeks vs 12 months). OS (HR, 1.16, 95% CI 1.03-1.31, P= 0.015,I2 0.00, p= 0.837) and DFS (HR, 1.14, 95% CI 1.02-1.26 , P = 0.016,I2 14.90, p= 0.319) were significantly worse in patients receiving shorter duration T as compared to 12 months. Cardiotoxicity was significantly higher in the 12 month group vs shorter duration (OR, 2.10, 95% CI 1.58-2.80, p = 0.000 ,I2 51.182, p= 0.085). Grade 3/4 cardiac events and cardiac events leading to discontinuation of therapy were significantly higher in patients receiving 12 months (HR 2.06, 95% CI 1.60-2.63, P = 0.000, I2 3.967, p= 0.384). There was no significant heterogeneity among studies. Risk of bias was low. Conclusions: Twelve months of adjuvant T is associated with significantly better DFS and OS compared to a shorter duration. As such, 12 months should remain the standard of care for most patients. There is an unmet need to identify lower risk groups which might do well with from shorter duration of HER2-directed therapy.

Introduction Chronic obstructive pulmonary disease (COPD) is a progressive lung disease, usually caused by tobacco smoking, but other important risk factors include exposures to combustion products of biomass fuels and environmental pollution. The introduction of several new (combination) inhaler therapies, increasing uncertainty about the role of inhaled corticosteroids and a rapid proliferation of the literature on management of stable COPD in general, call for novel ways of evidence synthesis in this area. A systematic review and evidence map can provide the basis for shared decision-making tools and help to establish a future research agenda. Methods and analysis This systematic review will follow an umbrella systematic review design (also called overview of reviews). We plan to conduct a comprehensive literature search of Ovid MEDLINE (including epub ahead of print, in process and other non-indexed citations), Ovid Embase, Ovid Cochrane Database of Systematic Reviews and Scopus from database inception to the present. We will include systematic reviews that assessed the effectiveness of any pharmacological or non-pharmacological intervention on one or more patient-important outcomes and/or lung function in patients with stable COPD. For every intervention/outcome pair, one systematic review will be included. An a priori protocol will guide, which systematic reviews will be chosen, how their credibility will be evaluated, and how the quality of the body of evidence will be rated. Data will be synthesised into an evidence map that will present a matrix that depicts each available treatment for stable COPD with a quantitative estimate on symptoms/outcomes from the patient perspective, along with an indication of the size and certainty in the evidence. Ethics and dissemination Approval by a research ethics committee is not required since the review will only include published data. The systematic review will be published in a peer-reviewed journal. PROSPERO registration number CRD42018095079

Background: Poly (ADP-ribose) polymerases (PARPs) are a highly conserved family of enzymes whose main function is to preserve genomic integrity following DNA damage. PARP inhibitors (PARPi) are increasingly used in cancers with deficiencies in homologous recombination. Clinical trials in breast and ovarian cancers have led to several FDA approvals in recent years, and their use in clinical practice is continuing to rise. It is thus necessary to assess their adverse event (AE) profile. Method: Literature search was performed using Ovid MEDLINE, EMBASE, CENTRAL, and Scopus (inception through October 26, 2018). Eligible studies were phase 3, randomized, controlled trials that compared single agent PARPi to placebo or standard treatment. Number of patients treated and AEs reported were recorded. Observed incidence of AE was reported with 95% CI. Heterogeneity was evaluated using Cochran Q statistic and I2 statistics quantified the proportion of heterogeneity not due to chance. Results: Databases revealed 869 references. Of these, 6 were eligible: 2 in breast cancer (OlympiAD, EMBRACA) and 4 in ovarian cancer (NOVA, ARIEL3, SOLO1, SOLO2). PARPi included niraparib (NOVA), olaparib (OlympiAD, SOLO1, SOLO2), rucaparib (ARIEL3), and talazoparib (EMBRACA). Of 1,685 patients who received PARPi, incidence of any AE, regardless of grade, was 98.5% (95% CI, 97.2–99.2%). Common AEs were: nausea (incidence rate, 68.9% and 95% CI, 58.7%–77.5%), fatigue (56.3%, 45.3%–66.8%), anemia (46.3%, 37.2%–55.8%), vomiting (33.7%, 29.5%–38.3%), neutropenia (24.7%, 15.3%–37.4%), headache (23.9%, 19.9%–28.4%), and reduced appetite (21.7%, 19.3%–24.3%). Myeloid neoplasms were rare (1.2%, 0.7%–1.9%). Incidence of grade 3 or higher AE was 44.3% (30.2%–59.5%) and often related to myelosuppression, specifically anemia (24.7%, 15.3%–37.4%), neutropenia (10.7%, 6.6%–16.9%), and thrombocytopenia (5.0%, 1.7%–14.0%). Incidence of serious AE was 24.3% (19.4%–29.9%); dose interruption occurred in 53.3% (41.2%–65.0%) and dose reduction occurred in 39.2% (23.6%–57.4%). 10% (7.4%–13.6%) of patients discontinued therapy due to AE. Death due to AE was rare; less than 1% (0.4%, 0.2%–0.8%) in all trials. Conclusion: Myelosuppression and gastrointestinal toxicities were the most commonly reported AE in 6 randomized phase 3 trials of PARPi for breast and ovarian cancers. Therapy was rarely discontinued due to AE. It remains to be seen whether these results will be reflected in clinical practice.

Abstract Background Osteoporosis and osteopenia are associated with increased fracture incidence in postmenopausal women. We aimed to determine the comparative effectiveness of various available pharmacological therapies. Methods We searched MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, ISI Web of Science, and Scopus for randomized controlled trials that enrolled postmenopausal women with primary osteoporosis and evaluated the risk of hip, vertebral, or nonvertebral fractures. A network meta-analysis was conducted using the multivariate random effects method. Results We included 107 trials (193,987 postmenopausal women; mean age, 66 years; 55% white; median follow-up, 28 months). A significant reduction in hip fractures was observed with romosozumab, alendronate, zoledronate, risedronate, denosumab, estrogen with progesterone, and calcium in combination with vitamin D. A significant reduction in nonvertebral fractures was observed with abaloparatide, romosozumab, denosumab, teriparatide, alendronate, risedronate, zoledronate, lasofoxifene, tibolone, estrogen with progesterone, and vitamin D. A significant reduction in vertebral fractures was observed with abaloparatide, teriparatide, parathyroid hormone 1-84, romosozumab, strontium ranelate, denosumab, zoledronate, risedronate, alendronate, ibandronate, raloxifene, bazedoxifene, lasofoxifene, estrogen with progesterone, tibolone, and calcitonin. Teriparatide, abaloparatide, denosumab, and romosozumab were associated with the highest relative risk reductions, whereas ibandronate and selective estrogen receptor modulators had lower efficacy. The evidence for the treatment of fractures with vitamin D and calcium remains limited despite numerous large trials. Conclusions This network meta-analysis provides comparative effective estimates for the various available treatments to reduce the risk of fragility fractures in postmenopausal women.

Background. This systematic review was commissioned to identify new variables associated with transplant outcomes that are not currently collected by the Organ Procurement and Transplantation Network (OPTN). Methods. We identified 81 unique studies including 1 193 410 patients with median follow-up of 36 months posttransplant, reporting 108 unique risk factors. Results. Most risk factors (104) were recipient related; few (4) were donor related. Most risk factors were judged to be practical and feasible to routinely collect. Relative association measures were small to moderate for most risk factors (ranging between 1.0 and 2.0). The strongest relative association measure for a heart transplant outcome with a risk factor was 8.6 (recipient with the previous Fontan operation), for a kidney transplant 2.8 (sickle cell nephropathy as primary cause of end-stage renal disease), for a liver transplant 14.3 (recipient serum ferritin >500 µg/L), and for a lung transplant 6.3 (Burkholderia cepacia complex infection for 1 y or less). OPTN may consider some of these 108 variables for future collection to enhance transplant research and clinical care. Conclusions. Evidence-based approaches can be used to determine variables collected in databases and registries. Several candidate variables have been identified for OPTN.